Introduction

It has been repeatedly shown on previous occasions that pleiomorphic and/or spindle shaped cells, immunoreactive for CD117/c-kit, were present in some malignant human gallbladder tumors (Ortiz-Hidalgo et al. 2000; Mendoza-Marin et al. 2002; Park et al. 2004; Langner et al. 2004; Furihata et al. 2005; for review see Thomas 2007). However, despite the fact that (in two case reports) it was speculated that the origin of such tumors might be Cajal-like cells (Ortiz-Hidalgo et al. 2000; Furihata et al. 2005), the presence of interstitial Cajal-like cell type (ICLC) in a human non-tumoral gallbladder has not yet been thoroughly documented. It appears obvious why for ethical reasons, literary accounts of human material have only been from patients with chronic cholecystitis (Hopwood and Ross 1997). Another line of evidence has been provided by a recent report about the presence of cells similar to interstitial cells of Cajal in CD1 mice (Sun et al. 2006) or guinea pig (Lavoie et al. 2007) gallbladders, respectively. The purpose of this study was to show, by means of a set of complementary techniques, the ‘expected’ presence, the distribution and ultrastructural characteristics, at high transmission electron microscopy (TEM) resolutions, of ICLC in human adult non-tumoral gallbladder. We applied a methodological algorithm previously used for identification and characterization of ICLC in some other non-enteric human and/or animal regions: pancreas (Popescu et al. 2005d), myometrium (Ciontea et al. 2005; Popescu et al. 2007), Fallopian tube (Popescu et al. 2005b), mammary gland (Popescu et al. 2005a; Radu et al. 2005), human atrial (Hinescu and Popescu 2005; Hinescu et al. 2006) and ventricular myocardium (Popescu et al. 2006a).

Materials and methods

Human tissue specimens

For immunohistochemistry, histological normal areas of gallbladder, remote of primary disease, were obtained from archived paraffin-embedded tissue of adult patients that had undergone surgical procedures for non-neoplastic disease. For a few cases, fresh gallbladder was used for vital staining. Archived paraffin and Epon-embedded specimens from a 17-week-old human fetus were processed to assess by immunohistochemistry and electron microscopy, respectively, the presence of gallbladder ICLC at an early stage of development.

Ethics

The study was conducted in accordance with the moral, ethical, regulatory and scientific principles governing clinical research as set out in the Declaration of Helsinki (1989) and Good Clinical Practice guidelines (GCP). Internationally recognized regulatory and ethical standards regarding ownership of tissue samples used for research were followed for archived tissue specimens (Hakimian and Korn 2004). This study was approved by the Bioethics Committees of the “Victor Babes” Institute of Pathology, Bucharest and “Carol Davila” University of Medicine.

Toluidine blue staining

Control semi-thin sections (less than 1 μm) were stained with 0.25% toluidine blue and examined by light microscopy (Nikon Eclipse E600). Representative photomicrographs were taken using Nikon Plan 40× and Nikon Plan Fluor 100×/1.30 oil.

Electron microscopy

A classical protocol was used to obtain Epon-embedded specimens. After the gallbladder was excised it was immersed 90 min, at 4°C, in a freshly prepared solution of 2.5% glutaraldehyde in 15 mM cacodylate buffer, pH 7.2. Tissue samples (fragments of about 1 mm3) were then fixed with 1% osmium tetroxide in 15 mM cacodylate buffer, pH 7.4, for 90 min at 4°C, and dehydrated in a series of graded ethanols. After immersion in propylene oxide (three times for 10 min each), the samples were immersed in a mixture (1:1) of propylene oxide and Epon 812 resin, overnight, and embedded in Epon 812. The regions studied were sectioned using a MT-7000 ultramicrotome (from Research Manufacturing Company Inc., Tucson, Arizona, USA) into ultrathin sections (50 nm). The sections were mounted on Formvar coated copper slot grids, stained with uranyl acetate and lead citrate, and then observed in a CM 12 Philips electron microscope at an acceleration voltage of 60 kV.

Morphometric analysis

The absolute and relative number of cells, as well as the number and length of processes emerging from the cell body were evaluated based on a combination of measurements of semithin and ultrathin sections obtained from 27 human gallbladders (from which 17 examined by TEM and 27 by light microscopy). Tissue specimens were from both male (n = 11) and female patients (n = 16), of 29 to 66-years-old. Two histopathologists (unaware of experimental design) counted the number of stained cells (excluding vascular structures) within photographs (n = 92) randomly taken on semi-thin sections. Similarly, 94 randomly taken TEM images were examined according to the ultrastructural diagnostic criteria for ICLC. Software for statistical analysis determined average values for different cellular types. Results were reported as mean plus or minus standard error.

Immunohistochemistry

The human gallbladder selected specimens were tested by immunocytochemistry for ICLC, by means of the following antibodies: CD117, polyclonal, 1:100 (DAKO, Glostrup, Denmark), CD34, 1:100, clone QBEND10 (Biogenex, San Ramon, CA, USA), chromogranin A, monoclonal, 1:50, clone LK2H10, (Novocastra, Newcastle upon Tyne, UK), smooth muscle α- actin (SMA), monoclonal, 1:1500, clone 1A4 (Sigma Chemical, St. Louis, MO), S-100, polyclonal, 1:500 (DAKO, Glostrup, Denmark), CD68, monoclonal, 1:50, clone PG-M1 (DAKO, Glostrup, Denmark), NK1, monoclonal, 1:50, clone NK1, (DAKO, Glostrup, Denmark), vimentin, monoclonal, 1:100, clone V9 (DAKO, Glostrup, Denmark), desmin, monoclonal, 1:50, clone D33 (DAKO, Glostrup, Denmark), tau protein, polyclonal, 1:100, (NeoMarker, Fremont, CA, USA), nestin, monoclonal, 1:100, clone 5326 (Santa Cruz, CA, USA), PGP9.5, monoclonal 1:40, clone 10A1 (Novocastra, Newcastle upon Tyne, UK), GFAP, monoclonal, 1:50, clone 6F2, (DAKO, Glostrup, Denmark), CD62P (selectin), monoclonal, 1:25, clone 1E3 (DAKO, Glostrup, Denmark) estrogen receptor, monoclonal, 1:25, clone 6F11, (Novocastra, Newcastle upon Tyne, UK), and progesteron receptor monoclonal, 1:50, clone 16 (Novocastra, Newcastle upon Tyne, UK). All specimens were counterstained with Mayer’s hematoxylin, examined and photographed on a Nikon Eclipse 600 microscope. Immunocytochemistry was performed on 3 μm-thick sections from 10% formalin fixed paraffin-embedded specimens, according to the Avidin-Biotin-Complex (ABC) method of Hsu (Hsu et al. 1981), modified (Bussolati and Gugliotta 1983) (for detailed preparative techniques see Miller 2002). Briefly, the procedure was: deparaffinization in xylene and alcohol series, rehydration, washing in phosphate saline buffer (PBS), incubation with normal serum, for 20 min, incubation with primary antibody overnight, standard labeled streptavidin–antibody biotin (LSAB) kit (DAKO), washing in carbonate buffer and development in 3-3′-DAB hydrochloride/H2O2. Microwave antigen retrieval (in M-citrate buffer, pH 6.0) was performed for CD117/c-kit, chromogranin A, S-100 and CD68. For CD34 and CD117 double immunostaining of the human gallbladder, an ABC indirect tristadial method was used: DAB (diaminobenzidine) chromogenic substrate developing for CD34 monoclonal antibody, and fast red alpha naphtol in Tris buffer for CD117 polyclonal antibody. Negative controls were maintained using an irrelevant primary antibody or replacing the secondary antibody with phosphate buffered saline. A positive control, a tissue known to express the marker in question, was used for every specific immunocytochemical stain. To ensure the reliability of this experimental study, internal quality control of immunocytochemical techniques was performed as a part of a certified quality assurance system.

Results

Methylene blue staining

By tradition, methylene blue staining is the first step in identifying cell profiles in a stroma (Cajal 1911). Figure 1A shows selectively-stained cells in fresh tissue samples in the human gallbladder. The characteristic features of ICLC (positioning in interstitium, networklike appearance, and thin, characteristic, very long, and moniliform cell prolongations) can be foreseen. Using the methylene blue vital staining followed by cryosectioning, we approximated a value of 100–110 cells/mm2. This value is in the same order of magnitude as previously reported for methylene blue staining in the fallopian tube (Popescu et al. 2005b). We estimated the spatial density of ICLC in adult human gallbladder and observed a dual distribution: in the close vicinity of the epithelium and in the interstitial spaces between bundles of smooth muscle fibers. Sometimes ‘beads’ of the moniliform cytoplasmic processes were observed.

Fig. 1
figure 1

Adult human gallbladder stained with methylene blue (A). Human gallbladder semithin sections stained with toluidine blue (non-conventional light microscopy). Interstitial cells having ICLC morphology, with thin cellular processes are illustrated. Cellular processes embrace crosssectioned (B) or longitudinally sectioned small diameter vessels (C). Long cellular processes running between smooth muscle bundles are suggestive for ICLC. Original magnification 40×. Red dot lines follow the ICLC profile in order to make it more evident

Full size image

Toluidine blue staining

This approach (consisting of tissue processing as for TEM, followed by sectioning around 1 μm and examination under a light microscope) allows a rough estimation of ICLC localization, quantity, or morphological details (e.g. cytoplasmic processes). In the human adult gallbladder, as shown in Fig. 1, ICLC are mainly placed near small vessels (Fig. 1B, C), in the subepithelial region of lamina propria and between smooth muscle bundles in muscularis. A combined analysis on semithin and ultrathin section enabled us to determine the relative proportion of ICLC in the sub-epithelial and muscularis interstitium (about 7%, and ∼5%, respectively). In toto, we estimated that gallbladder ICLC represent ∼5.5% of sub epithelial cells. Like in other organs (McCloskey et al. 2002; Bobryshev 2005; Ciontea et al. 2005; Lang and Klemm 2005; Popescu et al. 2005d; Radu et al. 2005; Harhun et al. 2005; Sergeant et al. 2006), ICLC cell shape is characteristic, with a long, thin process (several tens of μm).

Immunohistochemistry

In order further characterize ICLC, immunohistochemistry was used to correlate morphology with the cell phenotype. To this aim, 16 types of antibodies were assessed by single- or double-labeling IHC technique on formalin-fixed, paraffin-embedded sections, or cryosections from human adult lamina propria and muscularis of the gallbladder (Table 1). IHC using antibodies against CD117/c-kit revealed a high density of ICLC, mainly in the human adult gallbladder lamina propria (Fig. 2A) and mast cells. The positive control for CD117 immunoreactivity was performed on human myenteric plexus in the colon. CD117/c-kit expression was detected in cells with characteristic morphology susceptible to classification as ICLC (one or more very long, thin processes, sometimes with a beads-on-string appearance, on spindle-shaped cell bodies). In the adult human gallbladder lamina propria, ICLC constantly expressed CD34 at high levels (Fig. 2B). CD117/CD34 double immunohistochemical reaction was positive (Fig. 2C). CD117 mainly stains the cell body, while CD34 preferentially stains the cell processes. Almost a majority of ICLC, from lamina propria and muscularis, were non-reactive for α-SMA (Fig. 2D). In contrast, constant positive immunostaining was documented for vimentin (Fig. 2E), and nestin (Fig. 2G).

Table 1 Summary of immunohistochemical results for human gallbladder ICLC
Full size table
Fig. 2
figure 2

Immunocytochemical reactions for CD117 (A), CD 34 (B), double staining for CD117/CD34 (C), smooth muscle actin (D), vimentin (E), desmin (F), nestin (G), S-100 (H). Mayer’s hematoxilin counterstaining; original magnification A–E, G, H 40×; F 20×. Arrows indicate cells with a profile suggestive for ICLC

Full size image

Scattered cells were found positive for desmin (Fig. 2F). IHC with antibodies against some additional antigens showed inconstant or weak immunostaining for S-100 (Fig. 2H), and NSE (Fig. 3A), clear negative immunostaining for CD68 (Fig. 3B), GFAP (Fig. 3C), chromogranin A (Fig. 3D), PGP9.5 (Fig. 3E) and CD62 (Fig. 3G). In contrast, tau-protein was constantly positive (Fig. 3F). These data support the distinction between ICLC and other cells present in interstitium or adjacent layers (fibroblasts, pericytes, macrophages, mast cells, smooth muscle cells, neuroendocrine cells, neural cells). No significant differences were observed when immunohistochemistry in adult human gallbladder tissue specimens was compared to data on fetal human gallbladder.

Fig. 3
figure 3

Adult human gallbladder, immunostaining for: NSE (A), CD 68 (B), GFAP (C), chromogranin (D), PGP9.5 (E), tau protein (F), CD62P (G). Mayer’s hematoxilin counterstaining; Original magnification 40×

Full size image

Electron microscopy

TEM images revealed the presence of two subpopulations of gallbladder ICLC (located in the sub-epithelial region and between muscular bundles) with similar ultrastructural features: (1) Location in the non-epithelial space; (2) Close contacts with targets: nerve bundles, and/or epithelia, and/or smooth muscle cells, and/or capillaries; (3) Characteristic cytoplasmic processes: i. Number (1–3, frequently: 2–3); ii. Length (tens up to hundreds of μm); iii. Thickness (uneven caliber, <0.5 μm); iv. Aspect: moniliform, usually with mitochondria in dilations; v. Presence of ‘Ca2+ release units’; vi. Branching: dichotomous pattern; vii. Organization in network—labyrinthic system: overlapping cytoplasmic processes; (4) Gap junctions: with smooth muscle cells or with each other; (5) Basal lamina: occasionally present; (6) Caveolae: 2–3% of cytoplasmic volume; ∼0.5 caveolae/μm of cell membrane length; (7) Mitochondria: 5–10% of cytoplasmic volume; (8) Endoplasmic reticulum: about 1–2%, either smooth or rough; (9) Cytoskeleton: intermediate and thin filaments, as well as microtubules, present; (10) Myosin thick filaments: undetectable. Most of these specific features of ICLC are illustrated in Fig. 4. The inset shows the characteristic labyrinth of cell prolongations. Sometimes ultrastructural complex associations of caveolae, endoplasmic reticulum and/or mitochondria (putative as Ca2+ release units) were observed. Ultrastructure of human gallbladder interstitial cells in early stages of development (17 weeks) is presented in Fig. 5A, B. One may observe that early in development, a twisted system of cell prolongations is already formed by ICLC. Connections between ICLC and other connective cells have been previously reported in other tissues as stromal synapses (Popescu et al 2005c, 2006a). Figure 6A and B present a stromal synapse with a cleft width of 15.7 ± 2 nm, and wider inter-membrane distances (35.5–40.8 nm) observed in human adult gallbladder interstitium. At least in a few zones, it seems that the two membranes-contracting synapse are very closed together. Preliminary estimations revealed that there is an ultrastructural resemblance of gallbladder ICLC with both, interstitial cell of Cajal described in the musculature of gastrointestinal tract, and ICLC recently described in some other extra digestive organs. The only specific aspect of gallbladder ICLC seems to be the heightened twisting profile of cell processes.

Fig. 4
figure 4

Adult human gallbladder: TEM; original magnification: 40,000×. Photographic reconstruction illustrating the cell shape and a typical ICLC with long, thin, elaborate branching system of cell processes. Left image presents at a higher magnification area delimited by the thin border in the photographic reconstruction. N = nucleus; m = mitochondria; arrowheads = caveolae

Full size image
Fig. 5
figure 5

Fetal (17 weeks) human gallbladder: TEM; original magnification: 3,400× (A); 5,700× (B). (A) illustrates the location of gallbladder ICLC in the close vicinity of a vessel, and (B) illustrates the very early development of the long, labirinth-forming cell prolongations. N = nucleus; cap = capillary; arrows = ICLC cell prolongations; RBC = red blood cell

Full size image
Fig. 6
figure 6

Stromal synapse between an ICLC prolongation and another interstitial gallbladder cell (A). TEM, original magnification 10,000×. Contours of the two cell membranes in contact appear traced in (B). coll = collagen fibres

Full size image

Discussion

One can scrutinize the medical significance of the presence of ICLC in human gallbladder from different points of view: embryological, physiological and/or pathological.

From an embryological point of view, indirect evidence supports the idea that ICLC simultaneously develop in the intestine and bilio-pancreatic region. For the archetypal ICC, a kit receptor seems to be involved in the development and maintenance of the differentiated state of ICC after birth (Vannucchi 1999; Sanders and Ward 2006). As observed a long time ago, physiological disorders in the alimentary tract and unregulated secretion of bile are characteristic phenotypes of W mutant mice (Maeda et al. 1992; Isozaki et al. 1995). More recently, the fact that the biliary system, liver and pancreas might be considered as a unique organ was suggested (Frossard 2005). Molecular mechanisms which determine the organ identity in this complex area are not yet completely known (Sumazaki et al. 2004; Shen et al. 2002). The very close developmental relationships in this anatomical region might explain the transdifferentiation/plasticity of pancreas, biliary system and liver (Lardon et al. 2004) and provides a rationale to hypothesize that, if present in the pancreas (Popescu et al. 2005d), ICLC might be present in the gallbladder too.

From a histological point of view, data presented here represent a first methodical attempt to examine if, in non-tumoral human gallbladder interstitium cells corresponding to the morphological and immunocytochemical criteria for ICLC can be identified. For the first time a(n) (almost) complete immunohistochemical profile for both fetal and adult human gallbladder for ICLC is provided here. Several reports indicate that many gastrointestinal stromal tumours (GISTs) are positive for c-kit and CD34 and have other features similar to those of ICC (Streutker et al. 2007). In ICLC from the human gallbladder, we documented the concomitant presence of these two antigens. Recent studies on Cajal-like cells in CD1 mice gallbladder (Sun et al. 2006) and guinea pigs (Lavoie et al. 2007) assessed the exclusive presence of CD117, from an immunocytochemical perspective. Furthermore, one group initially reported the lack of c-kit positive cell in guinea pig gallbladder (Parr et al. 2003). However, it is worthwhile to mention that human and murine ICLC were similarly distributed, mainly in the sub epithelial and muscular layers. An overview of literature shows that, when attempts were made to culture in vitro gallbladder tissues, little attention was paid to interstitial cells (for details see Nakanuma et al. 1997).

From a physiological point of view, Cajal-type of interstitial cells might represent, through analogy to the gastrointestinal tract an essential player in the physiology of a digestive cavitary organ such as gallbladder, imposing the rhythm of bile release (pace-maker cells). It is noteworthy to observe that, when a comparison of phenotypic characteristics of c-kit mutant C57BL/6-Kit W-sh/W-sh and WBB6F1-Kit W/W-vmice was made, among other modifications, a lack of (digestive) ICC was concomitant with the exhibit of bile reflux, in both strains (Grimbaldeston et al. 2005). However, whether the physiology of ICC is species-dependent or whether ICC expresses similar mechanisms and functions in animals and humans remains to be established (Oldham-Ott and Gilloteaux 1997). Whether recent data obtained in CD1 mice gallbladder (Sun et al. 2006) or in guinea pig gallbladder (Lavoie et al. 2007) could be easily extrapolated for humans needs to be verified. On one hand, in some species, eating only periodically, as well as in humans, bile received from the liver becomes more concentrated, and is stored for longer time periods than the bile in vertebrates that eat frequently, such as guinea pigs (Oldham-Ott and Gilloteaux 1997). On the other hand, some other species, such as pigeons, rats, and deer, which eat almost continuously, have no gallbladder. Instead, in rat, for example, a bile ductular plexus seems to have storage and bile modification functions (Oldham-Ott and Gilloteaux 1997). The effects of imatinib on generation and propagation of action potentials, reported in guinea pig gallbladder preparations (Lavoie et al. 2007) also have to be verified in the human gallbladder and compared with results in other tissues (Popescu et al. 2006b). Further functional studies are required to confirm involvement of ICLC in pace-making and/or some other (paracrine, juxtacrine, autocrine) functions cannot be ruled out a priori (Rumessen et al. 2001; Cretoiu et al. 2006; Faussone-Pellegrini 2006; Faussone-Pellegrini and Vannucchi 2006; Faussone-Pellegrini et al. 2006; Ye et al. 2006). Starting from an immunostaining pattern similar to that of an intestitial cell of Cajal, a human gallbladder GIST-resembling tumor was considered as potentially deriving from mesenchymal stem cells. It implies that, in this case, putative ICLC (virtually regarded as stem cells) were considered as the starting point. The speculation went further, stating that stem cells residing at different sites could display dissimilar characteristics and stem cells in the gallbladder might possess the ability to differentiate more than in a classical GIST, even in a rhabdomyomatous direction (Furihata et al 2005) (previously not described for GIST; for review see (Bussolati 2005; Heinrich and Corless 2005; Robinson et al. 2000; Sanders et al. 2002).

From a pathological point of view, if one accepts a pace-making role, identification of Cajal-type cells in human gallbladder might enlarge the discussion about the role of this cell type in not only stromal tumor development (Ortiz-Hidalgo et al. 2000; Mendoza-Marin et al. 2002; Park et al. 2004; Langner et al. 2004, Furihata et al. 2005) but in human motor pathologies too. Depletion of ICC populations has been reported in tissues of patients with a variety of gastrointestinal motor disorders, including both congenital and acquired diseases (Sanders et al. 2002; Porcher et al. 2002; Lyford et al. 2002; Kubota et al. 2005; Taguchi et al. 2005, Tong et al. 2004). One might speculate that the same mechanism could be valid for gallbladder pathology. For instance, the reported depletion of ICC in patients with Crohn’s disease and the increased risk of pigment gallstone formation have apparently not been correlated. Similarly, it could be of interest to assess the gallbladder ICLC population in patients with home parenteral nutrition showing a high rate of cholelithiasis (Dray et al. 2007). In this context, as pure speculation, a (more) generalized decrease in the global ICC pool might be hypothesized as a new mechanism of disease, too. Recently, ICC depletion as a potential mechanism of disease has also been reported outside the digestive tract (Solari et al. 2003). By analogy with multiple endocrine neoplasia, one can imagine a ‘multiple-organ ICC/ICLC depletion’. This analogy overpasses the limits of simple assumption since, for example Hirschprung’s disease (an example involving ICC depletion) has in common with multiple endocrine neoplasia mutation(s) in RET tyrosine-kinase (for review see Takahashi 2001). As mentioned earlier, other indirect evidence of the presence of Cajal-type cells in the gallbladder was repeatedly provided in the last few years, when gallbladder tumors with a phenotype typical for interstitial cells of Cajal were reported (Ortiz-Hidalgo et al. 2000; Mendoza-Marin et al. 2002; Park et al. 2004; Furihata et al. 2005). An elegant study examined whether a microarray technique could be an appropriate tool to identify such tumors (and stated a negative conclusion) (Langner et al. 2004).

Conclusions

In this paper, we further substantiate the idea that the distribution of Cajal-type cells in humans might be on a larger scale than was originally thought (Huizinga and Faussone-Pellegrini 2005), documenting the presence of ICLC in fetal and adult non-tumoral human gallbladder interstitium. It was previously observed that heterogeneity of ICC, as well as organ or species specificity of corresponding types of ICC, have to be interpreted in a complex context, taking into account: origin (multipotency of a family of mesenchymal cells), microenvironment, mechanical force, nerve supply, relationships with muscle cells, or eating habits of animals, and movement patterns of the organ proper (Komuro 1999). Accumulating evidence suggests that this is the case for all types of ICLC as well. Cytological features exposed by ICLC depend upon the specific requirements of different tissues or organs. The gallbladder is not devoid of such peculiar requirements. Nevertheless, there is still a possibility of misdiagnosing ICC/ICLC (Streutker et al 2007). Extensive use of the available guides (Faussone-Pellegrini and Thuneberg 1999; Huizinga et al. 1997) and standards for identification for ICLC (Popescu et al. 2005b) should be encouraged in the future.